Concepts of network latency

 

Assume that A data packet X ranges from router A to router B. In this case, the total latency of data packet X from router A to router B is analyzed:

When A receives A packet, it detects the packet header and determines which link the packet is directed to. The time required for this process is called the node processing latency. The advantages and disadvantages of a router play a decisive role in processing latency.

After knowing which link the data packet is directed to, the next packet goes through the queuing latency. The queuing latency depends on the number of packets that are waiting to be transmitted to the link in queue. If the queue is empty and no other groups are being transmitted, the queue latency is 0.

When the data packet arrives in advance is transmitted, it passes through the queuing delay, and it is to push all data of X to the link. The time required for this process is the transmission delay. The transmission latency is determined by the length of data packet X and the link transmission rate. The unit of transmission speed of the link is bit/s. The transmission rate is mainly related to the speed of the Ethernet link. For example, in a 100 Mbit/s Ethernet link, the link transmission rate is 100 Mbit/s.

When all data of X is pushed to the link, the propagation delay is triggered. The propagation latency depends on the propagation distance and speed. The propagation speed depends on the physical media of the link, such as optical fiber cables and twisted pair wires. Generally, the propagation speed is equal to or less than the speed of light. The Unit is m/s.

After transmission latency, data packet X can reach the destination router B.

 

We can obtain a formula for calculating the total latency of a node:

Total node latency = node processing latency + queuing latency + transmission latency + propagation latency

 

This is an excerpt from the binlobby column.